AI Media Spend Reconciliation Agent

How Cassidy automates media spend reconciliation using AI

Step 1: Ingest and normalize media data

The Workflow triggers on new invoices, delivery reports, or IO updates—pulling data from ad platforms (Google Ads, Meta, DV360), ad servers (CM360, Sizmek), and your ERP. Cassidy normalizes units, rates, currencies, and vendor naming conventions into a canonical schema.

Step 2: Execute three-way match

Cassidy compares each IO line against delivery data and the corresponding invoice—checking quantities, rates, and amounts within your configured tolerances. Discrepancies are automatically categorized with root-cause suggestions (tag mapping errors, pacing variances, unauthorized fees).

Step 3: Route exceptions for resolution

Unmatched items are routed to exception queues with recommended remediation: makegood, credit memo, short-pay, or change order. Human-in-the-Loop approval ensures finance maintains control over business-critical decisions.

Step 4: Enforce sequential liability

Cassidy links campaign-level payables to client receivables, automatically releasing vendor payments only after client cash posts. The Workflow maintains a complete audit trail of the pay-when-paid chain.

Step 5: Generate accruals and close documentation

At period end, Cassidy calculates delivered-not-invoiced and prepaid accruals, generates journal entries with supporting schedules, and produces audit-ready packs—including variance explanations, approval trails, and client billing documentation.